Modelling comparison to evaluate the importance of phenology ... - Core

0 downloads 0 Views 1MB Size Report
Jun 22, 2018 - KEY WORDS. Climate change . ... by a structural rise in temperature were found: (1) a ... chemical photosynthesis model of Farquhar & von .... concentration alters both the initial light use efficiency ... symbols, with thelr units. ..... port-resistance model make it worthwhile validating ..... Differential equations.
Vol. 5: 119-130.1995

CLIMATE RESEARCH Clim Res

Published June 22

Modelling comparison to evaluate the importance of phenology for the effects of climate change on growth of temperate-zone deciduous trees Koen Kramer Institute f o r Forestry and Nature Research (IBN-DLO),PO Box 2 3 , 6 7 0 0 AA Wageningen, The Netherlands

ABSTRACT. The importance of 3 phenological types of deciduous trees for the effects of climate change on growth was evaluated using the model FORGRO The climate change scenarios used were a doubling of the CO2concentration (700 pm01 mol-') and an increase in temperature ranging from 0 to 7OC. To elucidate the relative importance of photosynthesis and allocation for this evaluation, models with different levels of mechanistic detail of photosynthesis and allocation were used. The photosynthesis approach of FORGRO was compared to the Farquhar & von Caemmerer approach as formulated in PGEN (FORGRO-PGEN).Similarly, the allocation approach of FORGRO was compared to the transport-resistance approach, as formulated in the ITE-Edinburgh model (ITE-FORGRO). A sensitivity analysis was performed to ascertain whether the response of gross photosynthesis to a climate change scenario depends on the value assigned to parameters in these models, and to compare this sensitivity with the differences found between the phenological types. The differences in the response of gross photosynthesis (P,) to the climate change scenarios between the phenological types were smaller according to ITE-FUKGKU as compared to FORGRO. These differences are of a similar magnitude when comparing the 2 photosynthesis models. Furthermore, FORGRO-PGEN showed that the response of P, to a 2x [CO,] Increases with nsing temperature, thus compensating for the Increase in respiration. For both FORGRO and ITE-FORGRO this CO2 and temperature interaction was not found. Consequently, in these models the increase in respiration exceeded the increase in gross photosynthesis at the higher range of temperature rise. The sensitivity analysis showed that the models differ in the sensitivity of the response of P, to a 2 x [CO2]scenario combined with a temperature rise of 2°C (C7ao/T2), when parameter values change by *25%. In FORGRO-PGEN, the magnitude of the response of P, depended on the values of some of its parameters, especially those determining the Michaelis-Menten kinetics of Rubisco, which for these parameters exceeded the differences between the phenological types in this scenario. In both FORGRO and ITE-FORGRO this sensitivity is similar to or less than the difference between the phenological types in the C700/TZ scenario.

KEY WORDS. Climate change . Deciduous trees Growth - Models . Phenology

INTRODUCTION As a result of natural selection, the annual biological cycle of the growth and dormancy of trees is synchronized to the annual climatic cycle of light, temperature and precipitation, thus determining growth. If the climate changes within the life-span of a tree, this synchronization may be partly lost. Consequently, either a part of the growing period of a tree may occur when the climate is not favourable for growth, or the growing period may not fully exploit the period when O Inter-Research 1995

the climate is favourable for growth. On the other hand, the species may be able to adjust by phenotypic plasticity. Earlier studies have predicted that, based on climate change scenarios, the probability of spring frost damage is likely to decrease in temperate zone Europe (Kramer 1994, Murray et al. 1989). It has also been found that trees do possess a considerable plasticity to accommodate a change in their local environment phenotypically (Kramer 1995). The aim of the study reported in this paper was to evaluate the importance of differences in phenological response to

Clim Res 5: 119-130, 1995

temperature for the effects of climate change on the growth of deciduous, temperate-zone tree species. Two models of photosynthesis and 2 models of allocation w e r e compared, to elucidate the consequences of describing these processes with different levels of mechanistic detail. In a n earlier study 3 phenological patterns induced by a structural rise in temperature were found: (1) a similar advance of both leaf unfolding and leaf fall; (2) a n advance of leaf unfolding, but no change in leaf fall; a n d (3) a larger advance of leaf fall than leaf unfolding (Kramer 1995).These 3 phenological types correspond to Betula, Fagus, a n d Quercus, respectively. Models incorporating detailed descriptions of light interception, photosynthesis, respiration and allocation are required to evaluate the effects of climate change on growth of deciduous trees. The models compared in this study were: (1) FORGRO (Mohren 1987, 1994) using the descriptions of photosynthesis of Goudriaan et al. (1985) and fixed keys for allocation; (2) FORGRO coupled to PGEN (Friend 1993), substituting the biochemical photosynthesis model of Farquhar & von Caemmerer (1982) for the photosynthesis model; and (3) FORGRO coupled to the ITE-Edinburgh model (Thornley 1991), in which the allocation keys of FORGRO a r e replaced by the transport-resistance approach of partitioning. Two aspects of climate change and growth of deciduous trees were studied through model companson: (1) the consequences of the phenological types on the effects of climate change scenarios on gross photosynthesis, a n d (2) the sensitivity of the scenario-induced response of gross photosynthesis to a change in parameter values of the models.

MATERIALS AND METHODS Phenology. To avoid inaccuracies in the date of both leaf unfolding and leaf fall in the analysis of the species response to the different scenarios, historical phenological observations for a 14 yr period were used. Phenological observations of Betula pubescens (birch), Fagus sylvatica (beech) and Quercus robur (oak) in The Netherlands were available for every year from 1940 untll 1953, except for 1945. For 1945 the average value of the phenological events was used. The phenological events monitored were leaf unfolding, full leaf a n d leaf fall. The observers had been provided with detailed instructions for each species, including pictures, of the exact event to observe, a n d instructions on how to select the trees (Anonymous 1950). The shifts of these events with either mean winter or summer temperature, based on a n extensive data set containing phenological observations of clones relocated over

a large latitudinal throughout Europe (Kramer 1995), are presented in Table 1. When the temperature was Increased according to a scenario, the observed dates of leaf unfolding, full leaf and leaf fall were adjusted according to the known responses of Betula, Fagus and Quercus (Table 1 ) . The shift in full leaf with winter temperature was assumed to be similar to leaf unfolding. Scenarios. Daily meteorological measurements for the period 1940 to 1953 were available for De Bilt (52" N, 6" E), located in the centre of The Netherlands, and used as input to the models. In all calculations, this series was adjusted according to a scenario. The vanable evaluated was the annual rate of gross photosynthesis, P, (t C H 2 0 ha-' yr-l), averaged over the simulation period. To evaluate the importance of phenology, the CO, concentration was set at 700 pm01 mol-l, and the temperature was increased uniformly by a maxlmum of 7°C in steps of 1°C. The benchmark scenario (no change in temperature) was also examined. The response for Pg of Betula, Fagus and Quercus to these scenarios was calculated according to the 3 models. The results were expressed relative to the scenario with [CO2] = 350 pm01 mol-l, without a n increase in temperature. The sensitivity of the response of P, to a change of ? 25% in parameter value was evaluated by comparing the response to the scenario with [CO2]= 700 pm01 mol-' and a uniform 2°C rise in temperature with the reference scenario with [CO,] = 350 pm01 mol-' and no increase in temperature. These scenarios will be referred to as C,,,/T, and C3,0/TO,respectively. The phenology of Betula (Table 1) was used for this analysis. Table 1 Phenological charactenstlcs of Betula, Fagus a n d Quercus. U average date of leaf unfolding; G . date of the stage full leaf; F: date of leaf fall; &U/ST,,,.change ~n date of leaf unfolding with mean winter temperature (T,,, 1 November unhl leaf unfolding); &GIST,,. change In date of full leaf; SF/ST, change in date of leaf fall w ~ t hmean summer temperature (T,, 1 May until leaf fall) I average c u m u l a t ~ v eirradiance from date of leaf unfolding to date of leaf fall, in The Netherlands (MJ m-2 growlng season-'), 6IU/ST,,.change in I caused by advancement of leaf unfolding (blJ "C 1, SIF/6T5: change In I caused by advancement of leaf fall (MJ " C - ' ) Bet~lla p -

U G F

I 6U/bT, SG/ST\, SIX/ST,, 6F/6T, 61r/ST,

22 Apnl 2 May 4 October 2504 -3 3 44(18%) -3 -24 (-1 0 % )

Fagus

Quercus

l May 8 May 16 October 2468 -2 -2 28(11%) 0 0 (0%)

5 May 15 May 20 October 24 13 -2 -2 32(13') -5 -28 (-1 1 %)

-

Kramer: Phenoloyy, tree growth and climate chanyc

Models. Three models with different levels of detail of photosynthesis and allocation were used, i.e. FORGRO, PGEN and the ITE-Edinburgh model. Briefly, FORGRO (Mohren 1987, 1994) is a process-based model suitable for predicting the growth of a n even-aged monoculture of coniferous tree species. The photosynthesislight response curve is modelled using a negative exponential function. An increase in the external CO2 concentration alters both the initial light use efficiency and the CO,-limited rate of gross photosynthesis (Goudriaan et al. 1985). Allocation of assimilates is modelled using fixed allocation keys. PGEN (Friend 1993) is a model aiming to predict the rate of photosynthesis at the biochemical level (Farquhar & von Caemmerer 1982), and the optimization of stomatal conductance given a set of environmental and biological parameters. The ITE-Edinburgh model (Thornley 1991) is a transport-resistance model of forest growth and partitioning based on counter-gradients of carbon and nitrogen substrate between foliage and roots. In the foregoing account the processes in which the models differ and those parts which were adjusted to calculate the growth of deciduous trees have been emphasized: see also Appendices 1 & 2

mamtenance carbohydrates growth resp~rat~on

+ -

121

Fig 2 Net CO, assimilation rate in relation to light absorption by the leaf surface. See Appendix 2 for an explanation of the symbols, with thelr units. (Figure redrawn from Goudriaan 1992)

FORGRO: Fig. 1 presents a simplified scheme of FORGRO. For photosynthesis, the minimum was taken of the rate of photosynthesis limited by either CO2 or the maximum value measured at light saturation (Figs. 2 & 3; Eqs. 1 to 4 in Appendix 2). Mesophyll resistance was calculated using: r, = (C, - T)/Fm,, (Fig. 2), assuming a constant ratio of internal to external CO2 concentration (Goudriaan et al. 1985). The boundary layer conductance was set a t a constant value, and the stomatal conductance depends solely on temperature. The temperature dependency of the CO2 compensation point is described using a mult~plier (Eq. 5). To relate the light-saturated rate of gross photosynthesis, a temperature n~ultiplierwas obtained b y !~Eo,=: :cterp=!atio:: =f 1itc:~turr data, using 6 hroai: plateau of near-unity in the range 10 to 30°C, and declining to zero outside this temperature range. A similar approach was taken to determine the actual mesophyll resistance a s a function of temperature, with values similar to the photosynthesis-temperature relationship. Daily gross canopy photosynthesis was calculated by integrating hourly over both sunlit and shaded leaf layers using a Gaussian integration scheme (Goudriaan 1986), dividing the canopy into 5 shaded and sunllt leaf layers. Growth and maintenance respiration were calculated using the approach of Penning d e Vries, which is based on the costs of biosynthetic processes and the biochemical composi-

litter

.......

matter

Fig. 1. Simplified diagram of the structure of FORGKO. Boxes: state variables; valves: rate variables; arrows: flows of carbon (solid lines) or infornlatlon (dotted lines). (Figure redrawn from Mohren 1994)

Fig. 3 Net C O 2 assimilation rate in relation to internal CO2 concentration See Appendix 2 for a n explanation of the symbols, wlth their units. (Figure redrawn from Goudriaan 1992)

Clim Res 5 : 119-130, 1995

irradiance, atmospheric pressure, air temperature and tion of the structural biomass (Penning de Vries et al. the absorbed photosynthetic active radiation at a given 1989). Fixed allocation keys were used for the growth rates of the different organs, with the exception of the leaf layer. Incidence of short wave radiation was set at twice the photosynthetic active radiation available at a allocation to the foliage and the reserve pool, for which saturation curves relative to maxlrnum values were given leaf layer. Output of PGEN is daily gross photosynthesis. used (Eqs. 6 & 7 ) . The level of reserves was modelled ITE-Edinburgh model: This model presents a mechusing a minimum equal to 5 % of the biomass of each anistic approach to assimilate partitioning based on organ, and a maximum which is 4 times as high. Allothe transport of labile carbon and nitrogen, and the cation of assimilates to the reserves has priority over all the other organs, once the full leaf stage has been size and activity of meristem (Fig. 4; Eqs. 23 to 29). The reached. Daily values of the meteorological variables transport of C and N substrate is driven by concentrairradiance, minimum and maximum temperatures, tion differences and resistances between the organs. humidity, wind speed and rainfall are required to run Counter-gradients of carbon and nitrogen substrate are formed because the foliage is the only source of C FORGRO, which uses a fixed time step of 1 d. PGEN: PGEN is a photosynthesis model which aims substrate, the roots are the only source of N substrate, at predicting stomatal conductance and photosynthesis and the growing organs act as sinks of carbon and with a minimal use of empirical parameterization. It is nitrogen. A functional root-shoot balance is attained based on the assumption that a leaf instantaneously because the acquisition of carbon depends on the level optimizes its stomatal conductance as a trade-off of N substrate of the foliage, and the acquisition of N between CO2 gain and water loss. CO2 gain affects depends on the level of C substrate in the fine roots. photosynthesis according to the biochemical photoThe growth of each organ is determined by the activity synthesis model of Farquhar & von Caemmerer (1982). and potential size of the meristem, which depends on The demand for CO2 is determined either by carboth the C and N substrate concentrations of the organ. Temperature dependency of parameters was boxylation limitation of Rubisco (Eq. g), or by regeneration limitation of RuBP (Eq. 10), while the supply of described using a parabolic-shaped multiplier, which CO2 depends on the difference of CO2 concentration equals zero at O°C, and is maximum at 30°C (Eq. 30). outside the leaf boundary layer and inside the leaf air The ITE-Edinburgh model was coupled to FORGRO spaces (Eq. 11). Whether the CO2 supply meets the (ITE-FORCRO) by using the modules of FORGRO photosynthetic demands depends on the resistance to CO2 along the pathLight interception and photosynthesis way from outside the leaf boundary layer to the mesophyll cells (Eq. l?). Explicit functions for r,,, and rc,iare presented in PGEN, while r,,, is the resistance which is optimized numerically. Eqs. (14) to (22) provide more detail on how the variables in Eqs. (9) to (12) are calculated. The leaf temperature is calculated from the leaf energy balance (Jones 1992). Temperature influences photosynthesis by altering the solubilities of CO2 and 02,and alters the MichaelisMenten constants of the carboxylation and oxygenation of Rubisco following the law of Arrhenius. The influence of Malntenance respiration temperature on dark respiration is GIOW~~ modelled by a Qlo approach. Carbon Nitrogen Structure. X Meridem. M substrale, N substrate, C PGEN was coupled to FORGRO by substituting it for the calculations of 'n'rlnslc dl"erentlalion Mer~slemact~vity requlres C and N substrates the gross photosynthesis (F,,,,,) in the Growth resp~ration canopy module and adjusting it so that CO? Litter and O2 'Oninput PGEN was: Fig. 4 . Simplified diagram of the structure of ITE-FORGRO. Light interception centration in the air, relative humidity, and photosynthesis are described as in FORGRO. (Figure redrawn from wind speed, incidence of short wave Thornley 1991)

F+=='

Kramer. Phenology, tree growth and cl~matechange

which calculate light interception, photosynthesis and stomata1 conductance. A reserve pool was required to start leaf growth after bud burst, and to allow for maintenance respiration in the leafless period. Therefore, a reserve pool was added for each organ. The growth rate of each reserve pool was set at a fixed fraction (0.05)of the growth rate of the structural biomass of the organ. Furthermore, it was assumed that the utilization of carbon and nitrogen and the respiration of the reserve pool are similar to the respiration of the structural biomass. During the build up of the canopy (the period from bud burst until full leaf), reserves are mobilized from all organs, i.e. converted into labile C and N, according to a first-order process. During this phase the foliage is the only organ allowed to grow. Consequently, a gradient of both C and N substrate from the fine roots to the foliage develops, since the foliage acts as the only sink. The leaves start to photosynthesize immediately, which causes the C substrate gradient to reverse as soon a s the carbon production exceeds carbon utilization, or when the full leaf stage is reached. During the leafless period, the costs of maintenance respiration are directly compensated for from the reserve pool of each organ. The leaf area index was truncated to the same maximum value as used in FORGRO. The ITE-FORGRO model was developed using SENECA v1.5, a Simulation ENvironment for ECological Application (De Hoop et al. 1992). The integration m ~ t h n dwa.~:Enlerian with variah!e t h e steps. ?reliminary runs indicated that it takes approximately 3 yr for the ITE-FORGRO model to attain stable gradients of labile carbon and nitrogen. Therefore, runs were started at 1937, using average values for the phenological events, but output of the 1940 to 1953 period is presented.

RESULTS Phenology

An in~pressionof the importance of the differences between the phenological types can be obtained by

123

examining the amount of light available on average during the growing period, and how this changes with a rise in temperature (Table 1 ) . On average, most irradiance is available for Betula. Fagus and Quei-cushave respectively 1.4 O/o and 3.6'%,less. When the temperature changes, the net result is a gain in the average available irradiance of 0.8% for Betula, 1.1% for Fagus and 0.2% for Quercus, per degree temperature rise, relative to the total cumulative irradiance available on average during the growing season for each of these phenological types. In The Netherlands, the irradiance gained on average when leaf unfolding is advanced by 1 d is more than twice what is lost when leaf fall advances 1 d (e.g. 15 MJ m-' d - ' on 1 May and 6 MJ m-' d-' on 15 October). Table 2 presents the results of FORGRO, FORGROPGEN and ITE-FORGRO for the C350/T0 scenario. Clearly, the differences in phenology only cause small differences in growth and radiation use efficiency, and are consistent with the pattern between the phenological types found in Table l. For this parameterization of the models the P, calculated by FORGRO is similar to ITE-FORGRO, but higher than that of FORGRO-PGEN. For this parameterization of ITEFORGRO, more carbon is respired by growth I-espiration than by maintenance respiration, whereas in FORGRO the opposite is true. Furthermore, the growth rates of the organs differ because of the different zechszism cf a!!ccat:cn (rcsu!ts not presented). The results of the 3 models when [CO2]= 700 pm01 mol-' are that differences in the response of P, between Betula, Fagus and Quercus increase with temperature (Figs. 5 to 7 ) . The difference in the response between Fagus and Quercus increases by approximately 4 % in the C7,),,/T2scenario and by approximately 20% in the C700/T7 scenario, for FORGRO and FORGRO-PGEN, but the corresponding increases according to ITE-FORGRO are 4 % and 13 %, because of the different mechanism of allocation. This is consistent with the differences between the phenological types based on the change in average available irradiance with temperature (Table 1). Figs. 5 to 7 further show that the response of P, to a doubled [CO2] is

Table 2. Results of FORGRO. FORGRO-PGEN and ITE-FORGRO for the C35n/T0scenario for the 1940 to 1953 situation using default parameter values. RUE: radiation use efficiency: ratio of annual total dry matter production and absorbed photosynthetically active radiation (g DM MJ-l). See Appendix 1 for explanation of the other symbols and their units

pg

Rm

4

RUE

Betula

FORGRO Fagus

Quercus

Betula

35.7 10.6 3.8 1.6

34.1 10.3 3.6 1.6

32.9 10.2 3.5 l .S

23.1 8.3 2.3 0.9

FORGRO-PGEN Fagus Quercus 22.2 8.0 2.2 0.9

22.1 8.0 2.2 0.9

Betula 33.1 3.7 7.5 1.5

ITE-FORGRO Fagus Quercus 31.8 3.6 7.2 l .5

31.5 3.6 7.1 1.4

Clim Res 5: 1 1 9 1 30, 1995

124

ITE - FORGRO

FORGRO

t Betula

-+

Fagus

-

-m-

Ouercus

Belula

-+

Fagus

+ Quercus

Fig. 5. Response of P, to 2 x [CO?]with increasing temperature, relative to the current climate ( Q ) ,according to FORGRO. Annual average over 1940 to 1953

Fig. 7. Response of P, to 2 x [CO2]with increasing temperature, relative to the current climate (Q),according to ITE-FORGRO. Annual average over 1940 to 1953

greatest according to FORGRO-PGEN, and least in ITE-FORGRO, and that the response increases with temperature according to FORGRO-PGEN (Fig. 6), but decreases with temperature according to both FORGRO a n d ITE-FORGRO (Figs. 5 & 7). T h e causes of the differences between FORGRO and FORGRO-PGEN are depicted in Figs. 8 & 9. For the current parameterization of FORGRO and FORGROPGEN it can be seen that: (1) FORGRO yields a higher P, than FORGRO-PGEN for any COz,temperature and light combination; (2) the sensitivity of P, to CO, at a constant light level increases with temperature according to FORGRO-PGEN, but decreases slightly according to FORGRO; (3)the sensitivity of P, to CO2 at 10°C increases with irradiance similarly in FORGRO and FORGRO-PGEN; and (4) there is a temperature and light interaction for the sensitivity of P, to CO2 accord-

ing to FORGRO-PGEN, but not according to FORGRO. The consequence of these differences between FORGRO and FORGRO-PGEN are that in FORGRO and thus ITE-FORGRO, the increase in respiration with temperature is not compensated for by a n increase in photosynthesis (Figs. 5 & ?), whereas this is the case in FORGRO-PGEN (Fig. 6).

Sensitivity analysis A sensitivity analysis was performed to evaluate which parameters are most important in determining the response of gross photosynthesis, P,, to an increase T

FORGRO

I

('C) ( ~ ~ m - ~ d - ' )

FORGRO-PGEN 1.6

0

1

2 -c Betula

3 4 dT ('C)

+ Fagus

5

6

7

-r Ouercus

Fig. 6. Response of P, to 2 x [CO,] w t h increasing temperature, relative to the current climate (Q), according to FORGROPGEN. Annual average over 1940 to 1953

0

100

200

300

400

500

600

700

external [CO2] (prnol mol-l)

Fig. 8. Response of P, to CO2 at d~fferenttemperature and light levels, according to FORGRO

Kramer: Phenology, tree growth and climate change

FORGRO-PGEN

external

[CO2l (pm01 rnoil)

Fig. 9. Response of P, to CO, at different temperature and light levels, according to FORGRO-PGEN

of both CO2and temperature. The response of Pg to the C700/T2scenario relative to the CSm/TDscenario was used to compare the sensitivities of the parameters. The general trend which can be seen for FORGRO is that when a parameter is set so that P, is lower than the default parameter value, then the response to the C7n"T2 scenario 1s greater fir^ 113) Fnr exzmp!e, a high ratio between internal and external CO2 concentration, Ci/Ca,reduces the P, relative to a low ratio; consequently P, is increased more by the C700/T2

20

D 10

0

FORGRO-PGEN

125

scenario compared with the low ratio ( 2 4 % versus 18%). High values of C,/Ca,the CO2 compensation point, and stomata1 resistance, and low values of the initial light use efficiency, the light extinction coefficient and specific leaf area reduce P,, and thus show However, for P,,,,, the the large response to C700/T2. opposite is true: the largest response to C7,0/T2is at the which clearly gives high values of high value of P,,, P,. This was caused by the fact that at low P,,,,,, this asymptote was met more frequently than at high P,,,, thus the sensitivity to the scenarios is less. In general it can be concluded that response of P, in FORGRO to the C,,,IT, scenario is similar over a wide range of values of the main parameters which determine light interception and photosynthesis. A clear effect of the PGEN formulation is that the response of Pg to the scenarios increases or decreases, depending on the value assigned to a parameter. This is especially true for the parameters describing the temperature response of a parameter (AS, m, n, E,, Ed). The reason for this can be seen from Eqs. (20)to (22):a change of 1 unit in a parameter in the exponent is equivalent to leaf temperature changing by approximately 0.03"C, because the temperature is presented in Kelvin. Thus, these parameters need to be estimated accurately, although a change of 25 % in the values of these parameters may exceed the range which is found experimentally. C$: the !TE-FORGRC mode!, the iiiojt pioiioiiiiccil effect was found for the total leaf nitrogen (NI,,,,,)and the fraction nitrogen in meristem and structural bioand fNiX). However, the magnimass of all organs (fNIM tude of the response of P, to the scenario is only slightly affected by a large change in the values of these parameters. The absolute response of the other parameters of the ITE-FORGRO model tested in this manner was much less than that of the nitrogen parameters, whilst only the coefficient determining the potential meristem size showed a P, response which differed more than 2 % between the scenarios.

-10

-20

Fig. 10. Difference (D) between the responses of P, to the C700/T2scenario relative to the benchmark scenario at high (+25%) and low (-25%) value of the parameter indicated

DISCUSSION AND CONCLUSIONS

Both FORGRO and FORGRO-PGEN showed that the difference in the response of gross photosynthesis to a doubled CO2 concentration between the phenological types ranges from 4 to 20 % if the corresponding temperature rises by 2 to ?"C,respectively. However, these models diverge in the degree of the

126

Clim Res 5 :

response of Py to doubled CO2 scenarios: in FORGRO this response ranges on average from +20% when there is no temperature rise to -16% when the rise is 7"C, while the corresponding range according to FORGRO-PGEN is +22% to +36%. These differences can be attributed to differences in the response of P, to [CO,]. In FORGRO-PGEN this response enhances when temperature and irradiance increase, whilst in FORGRO this interaction is weaker (Figs. 8 & 9). Consequently, in FORGRO-PGEN the increase in photosynthesis exceeds the increase in respiration, whereas in FORGRO and ITE-FORGRO the break-even point lies at or above a temperature increase of 5°C. The CO, X temperature interaction is frequently reported in the literature, and is stressed a s a n important aspect for the study of climate change effects (e.g. Idso & Idso 1994, Kirschbaum 1994). However, the absence of a response or a decline of the relative stimulation of biomass of perennial plants at high CO2 as temperature increases has also been reported (Ziska & Bunce 1994, a n d literature therein). According to the transport-resistance mechanism of allocation (Thornley 1991) the response of P, to the scenarios with doubled CO2 is less compared with FORGRO and FORGRO-PGEN: relative to the C3,,/T0 scenario it is + l 3 % for no temperature rise and -6% for a rise of 7°C. Callaway et al. (1994) presented experimental evidence for a reduced response of growth to enhanced CO2 because of an altered allocation pattern. They found that the initial stimulating effect of CO2 on the growth of Pinus ponderosa seedlings, and its enhancement by increased temperature, disappeared after 2 mo because of a n increased allocation of biomass to the roots and other non-photosynthesizing tissues. Furthermore, the differences in the response of P, to a 2x [CO,] scenario between the phenological types a r e less than FORGRO and FORGRO-PGEN: 4 % if the corresponding temperature rises by 2°C and 13% if it rises by 7°C (Fig. 6). These features of the transport-resistance model make it worthwhile validating this model for a number of tree species. Figs. 5 to 7 can be used to evaluate the temperature increase predicted by general circulation models (GCMs). Four well-known GCMs are OSU, GISS, GFDL and UKMO, which predict that mean annual temperature wlll increase by 3.0, 4.0, 5.3 and 6.5"C, respectively (Leemans 1992). However, these models use CO2 equivalents to calculate the increase in radiative forcing due to a n increase in greenhouse gasses. Approximately half of these greenhouse gasses is carbon dioxide, the other half consists of methane, CFCs etc. (Houghton et al. 1990).Furthermore, according to the GCM scenarios the temperature increases more during winter than during summer, rather than uniformly over the year (Leemans 1992). Conse-

quently, the GCM scenarios affect the timing of leaf unfolding more than the timing of leaf fall, and respiration during the growing season is less for the GChI scenarios than for the uniform temperature scenarios. Thus, the equivalent uniform temperature scenario involves a somewhat higher increase in temperature than the annual mean temperature increase of the GCM scenario. The sensitivity analysis of the parameters of the models affecting photosynthesis showed that for FORGRO and FORGRO-ITE there is generally little interaction between the value of a parameter and the degree of the response of growth to the C700/T2 climate change scenario, although many parameters strongly affect the response in absolute terms (Fig. 10). Typically, this sensitivity over a broad range of parameter values is similar in magnitude to the difference between the phenological types in the C700/T2scenario (Figs. 5 & 7 ) . For FORGRO-PGEN, however, the degree of the response of P, to the C7,,/T2 scenario depends on the value of a parameter (Fig. 10).This was especially the case for the parameters describing the Michaelis-Menten kinetics of Rubisco, and the effect of temperature on these parameters. Also the effect of nitrogen is such that at low values of the nitrogen parameters the response of P, to the C7,,/T2 scenario is greater than at high values of these parameters (Fig. 10). For these parameters, this sensitivity is greater than the difference between the phenological types in the C700/T2scenario (Fig. 6). The sensitivity of the response to a variation in the parameter values in FORGRO-PGEN indicates that these parameters must be determined accurately in order to evaluate the effects of CO2 and temperature on growth. Currently, they are available for only a few species. Furthermore, some of the parameters of the PGEN formulation vary considerably both between and within species (Wullschleger 1993). An analysis of uncertainty propagation in FORGRO sh.owed that variation in P,,,, QIo, and SLA within 95 % of their uncertainty limits yielded uncertainties of 19, 9, 9 and 2%, respectively, of the relative standard deviation of the annual growth rate (Van der Voet & Mohren 1994). In a sensitivity analysis of PGEN it was found that the sensitivity indices (ratio of the relative change in a parameter to the relative change in, net photosynthesis) of k,, K,,N,K,,E, ,,,, , ko, f ~ , r h land j,,, were 0.7, 0.6, 0.6, 0.4, 0.4, 0.3, 0.2 and 0.2, respectively (Friend 1995). Thus, the uncertainty or sensitivity of these output variables to a small variation in a parameter is not directly applicable for inferring the importance of this parameter on the effects of a climate change scenario on the output variable. In this study, only the direct effect of temperature on phenology was taken into account. However, nutrients

Kramer. Phenology, tree growth and chmate change

and C O 2 are known to interact with temperature. Murray et al. (1994) showed that for some Picea sitchensis clones, a n increased C O 2 yields a delayed bud burst and an advanced bud set under low nutrient supply. This could shorten the growlng season by 3 wk. Under high nutrient supply this effect was much less' Increasing temperature counteracted the 2" effect, resulting in a n advanced bud burst, which was less c o n ~ p a r e dto the situation where only temperature

was increased. Such complex interacting effects, which are clone specific, greatly complicate the evaluation of the effects of climate change on the growth of trees. Acknowledgements I thank Prof J . Goudriaan, Dr G h11 J I\dohren and the Ph D. group of the C T de \Ylt Graduate school of Production ~ c o i o g yfor valuable comments on earher drafts of the manuscript

Appendix 1. Symbols of vanables a n d parameters with the11dimensions. T h e value indicates the default value for the parameter Symbol

Definition

Units

Value

Variables pg

R,

4 G,

SLA

L,, P,,,,, l-,, Eo

C,/C, kd,, R,,, Qlo

Gross photosynthesis Maintenance respiration Growth respiration Growth: i = l , leaves; i = b, branches; 1 = S , stem; i = c , coarse roots; i = f , fine roots

Parameters Specific leaf area M a x ~ m u nleaf ~ area index Maximum rate of net photosynthes~s C O 2 compensation point at 20°C Initial light use efficiency Rat10 internal to external C O 2 concentration Light extinct~oncoefficient of canopy Dark respiration at 20°C Increase of Rd given 10°C temperature lncrease

t t t t

CH,O ha-' yr.' CH,O ha-' yr.' C H 2 0 ha-' yr.' DM ha-' yr-'

m2 k g - ' m2 (leaf) m-2 (ground) m g CO, m S - l pm01 mol-' kg CO, J-'

'

m g CO, m-,

S-'

20 6 0.56 50 0.45 0.7 0.65 0.028 2.0

Aljpa~~liix 2. Equations

FORGRO Leaf photosynthesis

Allocation I I,

F

=

F , l / l - e ' , #>)-R,,

(1)

a,,

=

R - R,,, R,,,

a,,

=

l -(a, +a,)

\

F, ,

M

8 8

"

C

"

C , -l- -

F",c

:

r,,

+ 1 6r, + 1.4r,

C, - lC, + 21-

E

=

Eo-

l-

=

l-20e00'(T-20)

X

+

d

(2) (3)

(4)

(5)

a,,, a,, a b , Allocation of asslrmlates to the asr reserve pool, leaves, branches a n d stem c, Ambient C O 2 concentration pm01 mol-l F,,,,, Max, gross photosynthesis m g CO, m-* S-' F" Net rate of photosynthesis m g CO2 m-, S-' F",c CO2 limited net photosynthesis m g CO, m-2 S-' F,,,,,,, Maxlmum net photosynthesis at high CO, a n d light levels m g CO, m-, S-' F Maximum endogenous photosynthetic capacity at high C O 2 m g CO, m-2 ss' and Light levels Habs Absorbed PAR J m-2 S-'

R, R,,,

Reserve pool, a n d maxlmum level of reserve pool k g CH,O ha-' Rd Dark respiration rate mg CO2 m-2 S-' r,, r,, rb Mesophyll, stomata1 a n d boundary layer resistance S m-' 7 Temperature "C L, L,,, Leaf a r e a index, and maxim2 (leaf) m-' (ground) m u m leaf area index l-, l-,, C O 2 c o m p e n s a t ~ o npoint, a n d CO2 compensation point at 20°C pm01 mol-' E Initial l ~ g h use t efficiency pg C O 2 J-' (Append~x cont~nuedon next page)

C h Res 5: 119-130, 1995

28

Appendix 2 (continued) PGEN

Temperature functions

Leaf photosynthesis

=

4 . 1

.-I

=

J ( C ,-r.) 4.5 C, + 10.5I-.

0.5 V,,,,,

- Rd

(10)

S,

K,0,

A ,,

An,,

mol

PAR-saturated electron transport rate k,, k, Rublsco carboxylation, and oxygenation turnover number M-M constant for carboxylation, K,, K, and oxvqenation of Rubisco (air space equivalents) K,,rhl hZ-M constant for carboxylation, K and oxygenation of Rubisco (temperature dependent)

m P -, i: c, o L . . PO

e RT

v,,,,, K,

Carboxylation-limted, RuBP regeneration-Limited, and stomata1 resistance-l~m~ted rate of net photosynthes~s m01 CO2 m-2 SS' [CO2]in leaf air spaces, and in air C,. C, outside the leaf boundary layer mol m-3 Concentration of air in leaf D internal air spaces m01 m-" E Transpiration mol H 2 0 m-2 S-' E, Activation energy J mol-l Ed Deactivation energy J mol-l Et Leaf Rubisco catalytic slte content in leaf m01 m-* fs,rub, fschl Fraction nitrogen In Rub~sco, and chlorophyll Hdbs Absorbed PAR quanta m - S-' J Potential electron transport rate mol e- m-2 ss' lmax PAR-saturated potential electron transport rate m.01 e- mol chl-l S-' (temperature dependent)

A ,,

=

i

l l

Jmdx

m01 e- m-2 S - ' mol (m01 s~tes).'S-'

mol m-3 m01 m-3

I

Leaf nitrogen content kg m-2 0, concentration in leaf air spaces mol O2 m-3 Atmosphenc pressure, and standard atmospheric pressure Pa Gas constant J K-' mol-' Resistance to CO2 from air outside the leaf boundary layer S m-' to the mesophyll surface Resistance to CO2 transfer S m-' across leaf boundary layer Reslstance to CO2 from Inside leaf surface to mesophyll surface S m-' Reslstance to CO2 across leaf surface S m-' m01 CO2 m-2 S-' Mitochondrial respiration Mitochondrial respiration mol CO2kg Y - ' S-' (temperature dependent) Solubility ot CO2,and O2 in water mol m-3 Average of leaf and air temperature, and leaf temperature K Maxlmum rate of carboxylation, m01 CO2 m-2 S" and oxygenation of Rubisco Photosynthesis compensation [CO2]in leaf air spaces in absence m01 CO: m-3 of rnitochondrial respiration J K-' mol-' Entropy parameter Empirical constants

Kramer: Phenology, tree growth and climate change

Appendix 2 (continued) ITE-Edinburgh Temperature function

Differential equations

Leaves (1): dM,c dt

-

TC,,+IA~ - Tc,rl,-l,

- R,,,

-%G

+

Mr.,

Fine roots (f):

Growth rate of meristem Growth rate of reserves Growth rate of structure Loss in merisiems ro intrinsic differentiation Loss In reserves to litter Loss in structure to litter Mobilization of carbon from reserves Mobilization of nitrogen from reserves Canopy gross photosynthesis rate

LITERATURE CITED Anonymous (1950) Manual to conduct phenological observations. Royal Dutch Meteorological Institute, Study-circle for Ecology and Phenology, Illb, model 2006 (in Dutch) Callaway RM, DeLucia EH, Thonlas EM, Schlesinger WH (1994) Compensatory responses of CO2 exchange and biomass allocation and their effects on the relative growth rate of ponderosa pine on different CO2 and temperature regimes. Oecologia 98:159-166 De Hoop BJ, Herman PMJ, Scholten H, Soetaert K (1992) SENECA 1.5 A Simulation ENvironment for Ecological Application. Netherlands Institute of Ecology - Centre for Estuarine and Coastal Ecology, Yerseke, ECOLMOD Report EM-6, p 180 Farquhar GD, von Caemmerer S (1982) Modelling of photosynthetic response to environmental conditions. In: Lange OL, Nobel PS, Osmond CB, Ziegler H (eds) Physiological plant ecology 11: Water relations and carbon assimilation, 12B. Springer-Verlag, Berlin, p 549-587

RI,

Tcir, T ~ i ~ j

L , T,.T,,

UN

Maintenance respiration Carbon transport flux Nitrogen transport flux 1,,I 'I'emperature; minimum, maximum and reference temperature Utilization of carbon for growth Utilization of nitrogen for growth Uptake of nitrogen from soil

kg dm ha-' d-: kg C ha-' d-' kg C ha-'

"C

kg N ha-' d-' kg N ha-' d"

Friend AD (1993) Use of a model of photosynthesis and leaf microenvironment to predict optimal stomata1 conductance and leaf nitrogen partitioning. Plant Cell Environ 14: 895-905 Friend AD (1995) PGEN: a n integrated model of leaf photosynthesis, transpiration, and conductance. Ecol Model 77: 233-255 Goudriaan J (1986) A simple and fast numerical method for the computation of daily totals of crop photosynthesis. Agr Forest Meteor01 38:249-254 Goudriaan J (1992) Simulation of crop growth. Report nr F300-201, Department of Theoretical Production Ecology, Agricultural University Wageningen Goudriaan J , van Laar HH, van Keulen H, Louwerse W (1985) Photosynthesis, CO, and plant production In: NATO AS1 Series. Series A. Wheat growth and model~ng,Life Sciences 86, p 107-122 Houghton JT, Jenkins GJ. Ephraums JJ (1990) Climate change. The IPCC scientific assessment. Cambridge University Press, Cambridge, p 364

Clim Res 5: 119-130, 1995

Idso KE, Idso SB (1994) Plant responses to atmospheric CO, enrichment in the face of envlronmental constraints: a review of the past 10 years' research. Agr Forest Meteor01 69:153-203 Jones HG (1992) Plants and microclimate. A quantitative approach to envlronmental plant phys~ology.Cambridge University Press, Cambridge, p 428 Kirschbaum IMUF (1994) The sensitivity of C, photosynthes~s to increasing CO, concentration: a theoretical analysis of its dependence on temperature and background CO2 concentration. Plant Cell Environ 17:?4?-754 Kramer K (1994) A modelling analysis on the effects of climatic warming on the probability of spring frost damage to tree species in The Netherlands and Germany. Plant Cell Environ 17:367-377 Kramer K (1995) Phenotypic plasticity of the phenology of seven European tree species, in relation to climatic warming. Plant Cell Environ 18:93-104 Leemans R (1992)Modeling ecological and agricultural impacts of global change on a global scale. J Sci Ind Res 51:709-724 Mohren GMJ (1987) Simulation of forest growth, applied to Douglas fir stands in the Netherlands Thesis, Agricultural University Wageningen Mohren GMJ (1994) Modelling Norway spruce growth In relation to site conditions and atmospheric COz. In: Veroustraete F. Ceulemans R (eds) Vegetation, modelling and

climate change effects. SPB Academic Publishing bv, The Hague, p 7-22 Murray MB, Cannell MGR, Smith RI (1989) Date of budburst of fifteen tree species in 13ritain following climatic warming. J appl Ecol26:693-700 Murray MB, Smith RI, Leith ID, Fowler D, Lee HSJ, Friend AD, Jarvis PG (1994) Effects of elevated CO2, nutritlon and climatlc warming on bud phenology in Sitka spruce (Picea sitchensis) and their impact on the risk of frost damage. Tree Physiol 14:691-706 Penning de Vries FWT, Jansen DM. Ten Berge HFM. Bakema A (1989) Simulation of ecophysiological processes of growth in several annual crops. Simulation Monographs, 29, PUDOC, Wageningen Thornley JHh4 (1991) A transport-resistance model of forest growth and partitioning. Ann Bot 68:211-226 Van der Voet H, Mohren GMJ (1994)An uncertainty analysis of the process-based growth model FORGRO. Forest Ecol Manage 691157-166 Wullschleger SD (1993) Biochemical limitations to carbon assimilation in Cg plants - a retrospective analysis of the A/C, curves from 109 species. J exp Bot 44:907-920 Ziska LH, Bunce J A (1994) increasing growih iemperdiule reduces the stimulatory effect of elevated CO2 on photosynthesis or biomass in two perennial species. Physiol Plant 91:183-190

Editor: G. Esser, GieBen, Germany

Manuscript first received: November 2, 1994 Revised version accepted: March 2% 1995